Admin 3julmthh: Verified competitive analysis: who does what, where they stop, where AIDAVA fits, honest gaps

2026-04-14T17:46:45Z

Verified competitive analysis: who does what, where they stop, where AIDAVA fits, honest gaps

New page

__NOTOC__
Competitive landscape analysis for [[AIDAVA]]. Which companies work on similar problems, what they do, where they stop, and where AIDAVA fits.

All data sourced from company websites, EU-Startups, TechCrunch, Crunchbase, proff.no, allabolag.se, and academic publications. See individual company pages for source links.

== Market Map: Who Does What ==

Health data curation is a pipeline with multiple steps. No single existing commercial product covers the full pipeline from raw heterogeneous data to published, interoperable, reuse-ready health records.

{| class="wikitable"
|-
! Pipeline Step !! What It Means !! Who Does It !! AIDAVA?
|-
| '''1. Data Collection''' || Gathering raw records from hospitals, labs, EHRs || [[PicnicHealth]] (US, patient-anchored), [[1upHealth]] (FHIR-based), [[TriNetX]] (site-based network) || Yes — ingests structured + unstructured data
|-
| '''2. Data Standards''' || Structuring data using FHIR, HL7, openEHR || [[Better]] (FHIR platform), [[InterSystems]] (HealthShare), [[Lifen]] (document exchange) || Yes — FHIR resource profiles as structural framework
|-
| '''3. NLP / Text Extraction''' || Extracting structured data from clinical narratives, discharge notes, reports || [[Cogstack]] (NHS NLP), [[Qantev]] (claims NLP), [[Averbis]] (German NLP) || Yes — deep learning for 3 languages
|-
| '''4. Knowledge Graphs''' || Representing clinical data as semantic graphs with ontology relationships || [[Healx]] (drug repurposing KGs), [[Ada Health]] (medical knowledge base) || Yes — PHKG based on SNOMED/LOINC/FHIR ontologies
|-
| '''5. FAIRification''' || Making data Findable, Accessible, Interoperable, Reusable || [[Castor EDC]] (trial data FAIRification), academic projects (Swiss PHN) || Yes — automated FAIRification pillar
|-
| '''6. Patient Interaction''' || Patients see, understand, and approve curation decisions || [[PicnicHealth]] (record access), [[Loovi]] (biomarker results) || Yes — explainable AI adapted to user knowledge level
|-
| '''7. Multi-stakeholder Reuse''' || Same curated data serves patients, clinicians, AND researchers || '''No existing commercial product does this''' || Yes — "curate once, reuse many times"
|}

== Verified Competitor Capabilities ==

=== Data Collection Layer ===

'''[[PicnicHealth]]''' (US, founded 2014, $60M+ raised)
* '''Verified:''' Collects medical records from US hospitals via single HIPAA authorization. Connected to 10,000+ healthcare facilities. 12 of top 20 pharma customers. 98% patient retention.
* '''What it does NOT do:''' Does not FAIRify data. Does not create knowledge graphs. Does not publish reusable data across stakeholders. US-only.
* '''Source:''' picnichealth.com, TechCrunch ($25M round), Fierce Biotech ($60M round)

'''[[1upHealth]]''' (US, founded 2017, $40M Series C)
* '''Verified:''' FHIR-based patient data aggregation platform. CMS Patient Access API compliant. $40M Series C led by Sixth Street Growth (2023).
* '''What it does NOT do:''' Does not curate or FAIRify data. Does not do NLP. Does not create knowledge graphs. Platform for data access, not data quality.
* '''Source:''' 1up.health, Sixth Street press release

=== Data Standards Layer ===

'''[[Better]]''' (Slovenia, founded 2016, €10M+)
* '''Verified:''' Open-source FHIR platform used by health systems across Europe.
* '''What it does NOT do:''' Does not do AI curation. Does not do NLP. Does not create knowledge graphs. Provides the standard, not the intelligence.
* '''Source:''' better.care

'''[[InterSystems]]''' (US, founded 1978, ~$1B+ revenue, ~1,800 employees)
* '''Verified:''' HealthShare platform in 100+ countries. FHIR-based data integration. Private company, self-funded. Launched HealthShare AI Assistant.
* '''What it does NOT do:''' Enterprise infrastructure — does not automate curation, does not do patient-facing explainability, does not create PHKGs.
* '''Source:''' intersystems.com, Healthcare IT Today (Top 100 ranking)

=== AI/NLP Layer ===

'''[[Cogstack]]''' (UK, founded 2014, open-source)
* '''Verified:''' Open-source NLP platform for extracting structured data from unstructured NHS clinical text. Developed at UCL and Guy's and St Thomas' NHS Foundation Trust.
* '''What it does NOT do:''' Does not create knowledge graphs. Does not FAIRify. Does not provide patient interaction. NHS-specific deployment.
* '''Source:''' cogstack.org, Nature Digital Medicine (clinical NLP survey)

'''[[Healx]]''' (UK, founded 2014, $47M+ raised)
* '''Verified:''' Uses knowledge graphs and AI for rare disease drug repurposing. $47M raised. Published academic research on knowledge graph approaches.
* '''What it does NOT do:''' Knowledge graphs are for drug discovery, not patient record curation. Does not handle patient data, does not do FAIRification, no patient interaction.
* '''Source:''' TechCrunch ($47M round)

'''[[Qantev]]''' (France, founded 2019, €30M raised 2025)
* '''Verified:''' AI-driven claims platform for health insurers. €30M raised (2025). Uses small AI models that outperform LLMs for health data processing.
* '''What it does NOT do:''' Focused on insurance claims processing, not patient record curation or knowledge graphs.
* '''Source:''' EU-Startups, TechCrunch

=== Federated Learning ===

'''[[Owkin]]''' (France, founded 2016, $300M+ raised)
* '''Verified:''' Federated learning for health data — trains AI models across hospital data without moving patient records. $300M+ raised. Partnership with Sanofi. Published Nature Medicine research.
* '''What it does NOT do:''' Does not curate individual patient records. Does not create PHKGs. Does not do FAIRification. No patient-facing interface. Focus is on model training, not data quality.
* '''Source:''' owkin.com, Nature Medicine (federated learning paper)

=== Research Networks ===

'''[[TriNetX]]''' (US, founded 2013, $40M+)
* '''Verified:''' 300+ healthcare organizations across 30 countries. Federated clinical data network for trial optimization.
* '''What it does NOT do:''' Queries existing data — does not curate, FAIRify, or create knowledge graphs. Site-based, not patient-anchored.
* '''Source:''' trinetx.com

'''[[Castor EDC]]''' (Netherlands, founded 2012, €25M+)
* '''Verified:''' Electronic Data Capture used in 10,000+ clinical studies. Published research on de-novo FAIRification via EDC.
* '''What it does NOT do:''' Captures trial data — does not curate heterogeneous hospital data. Does not do NLP or create knowledge graphs.
* '''Source:''' castoredc.com, academic FAIRification paper

== Where AIDAVA Fits ==

AIDAVA is the only project that combines ALL seven pipeline steps in one orchestrated system:

{| class="wikitable"
|-
! Capability !! AIDAVA !! Nearest Competitor
|-
| Collects heterogeneous data (structured + unstructured) || ✅ Yes || PicnicHealth (but US only, no curation)
|-
| NLP extraction from narrative text in multiple languages || ✅ Dutch, German, Estonian || Cogstack (English only, NHS only)
|-
| Creates Personal Health Knowledge Graphs || ✅ PHKG per patient || Healx (but for drugs, not patients)
|-
| Automated FAIRification || ✅ Yes || Castor EDC (but for trials only)
|-
| Patient-facing explainable AI || ✅ Adapts to user knowledge || No competitor does this
|-
| Multi-stakeholder reuse (patients + clinicians + researchers) || ✅ "Curate once, reuse many" || No competitor does this
|-
| Cross-language (same ontology, multiple languages) || ✅ 3 languages || No competitor does this
|}

== Verified AIDAVA Facts ==

From AIDAVA's own documentation and CORDIS:
* Grant 101057062, EUR 7.7M, 48 months (Sep 2022 — Aug 2026)
* 12 European partners + 2 associated partners
* Coordinator: Maastricht University (Remzi Celebi)
* Two use cases: breast cancer registries, cardiovascular longitudinal records
* Three test languages: Dutch, German, Estonian
* Tested at 3 university hospitals with emerging personal data intermediaries
* Technology: AI virtual assistant with backend curation tools + frontend human-AI interaction

'''Source:''' CORDIS Grant 101057062, aidava.eu, AIDAVA wiki page

== What AIDAVA Does NOT Have (Honest Gaps) ==

* '''No commercial product yet''' — research project ending Aug 2026
* '''No funding beyond EU grant''' — EUR 7.7M vs PicnicHealth ($60M), Owkin ($300M), InterSystems ($1B+ revenue)
* '''No paying customers''' — all competitors have commercial traction
* '''Limited deployment scope''' — 3 hospitals in 3 countries vs PicnicHealth (10K+ facilities), TriNetX (300+ organizations)
* '''No patient app in production''' — competitors have live products
* '''Research-to-commercialization risk''' — many EU research projects don't translate to products

== Competitive Risks ==

Companies that could add AIDAVA's missing pieces:
* '''PicnicHealth''' — could add FAIRification and knowledge graphs (has $60M and pharma customers)
* '''Better''' — could add AI curation layer on top of FHIR
* '''Owkin''' — could add patient-facing features (has $300M)
* '''InterSystems''' — could add automation and explainability (has $1B+ revenue)
* '''1upHealth''' — could add curation on top of FHIR data access

AIDAVA's advantage is the RESEARCH DEPTH: the ontological architecture, multi-language NLP, and reference knowledge graph. But translating this into a defensible commercial product before competitors close the gap is the key challenge.

== See Also ==
* [[AIDAVA]]
* [[AIDAVA Related Companies]]
* [[Ontologies for Longitudinal Health Records]]
* [[Companies]]

[[Category:Topics]]

AIDAVA Competitive Analysis - Revision history

Admin 3julmthh: Verified competitive analysis: who does what, where they stop, where AIDAVA fits, honest gaps